Evolution of the translator inPeriplocaceae andAsclepiadaceae

This paper deals with the problem of homology between the translators inPeriplocaceae andAsclepiadaceae. Since most differences between the flowers in both families are based on the functional differences in the pollination process, it seems that the question of common ancestry can best be solved by clarifying the problem of the evolution of the translator itself. Ontogenetic investigation has revealed that the central element of the periplocoid translator is the adhesive disc together with the base of the stipe. The translator inSecamoneae (the most primitive tribe inAsclepiadaceae) is homologous to this element, and inSecamone the adhesive disc is transformed into the corpusculum while the stipe is represented by the dorsal process to which the pollinia are attached. The translator in the genusFockea represents an intermediate stage in the further phylogenetic elaboration of the asclepiad translator. It corresponds toSecamone in the development of a dorsal process for the attachment of the pollinia and in the special mode by which the corpusculum is attached to the anther wings. On the other hand, the elongated shape and the formation of a floor in the corpusculum are derived traits pointing to the more highly evolvedAsclepiadaceae, such as theAsclepiadeae. The question of the derivation of the caudicles is discussed and a hypothesis concerning their development from pre-adaptive elements inFockea is put forward. General conclusions are 1. thatPeriplocaceae andAsclepiadaceae are sister groups and that the common ancestor possessed a primitive translator composed of at least the two elements adhesive disc and stipe; 2. that the morphological difference of the translators betweenPeriplocaceae andSecamone is smaller than betweenSecamone and the more advancedAsclepiadaceae.

---

Zusammenfassung Ein wesentlicher Unterschied zwischen den beiden FamilienPeriplocaceae undAsclepiadaceae besteht in der Ausbildung der Translatoren, mit deren Hilfe der Pollen von Insekten übertragen wird. Bei denPeriplocaceae kleben die Pollentetraden an dem Translator, dieser selbst haftet mit einer basalen Klebscheibe am Kopf des Pollinators. DieAsclepiadaceae verzichten demgegenüber auf den Klebstoff, die Pollinien sind durch zwei Caudiculae mit einem Klemmkörper verbunden, der am Insekt klemmt. Im Zentrum dieser Arbeit steht die Frage, ob sich zwischen beiden Translatorformen Homologien finden lassen, die eine gemeinsame stammesgeschichtliche Herkunft begründen.Der Translator derPeriplocaceae besteht aus den drei Elementen Klebscheibe (adhesive disc), Stipes (stipe) und Pollenschaufel (spoon). Im typischen Fall sitzt der Stipes dorsal der Klebscheibe an, wodurch die Klebfläche selbst wie ein Stempel nach unten bzw. schräg nach außen gerichtet ist. Demgegenüber ist die geradlinige Anordnung von Stipes und Klebscheibe beiPeriploca graeca eine Ausnahme, die nicht als Basis für den Vergleich mit dem Asclepiadaceen-Translator dienen kann. In der Ontogenese wird der feste Teil der Klebscheibe, das Scutellum, zuerst angelegt und durch basale Sekretaddition als etwa rechtwinklig zur Griffelkopfoberfläche stehende Scheibe hinausgeschoben. Die Bildung des Scutellums erfolgt also von seiner Schmalseite aus. Im nächsten Schritt wird der Stipes-Ansatz am Scutellumrücken gebildet, auf diese Weise entsteht die gewinkelte Verbindung von Klebscheibe und Stipes. Der Anschluß der Pollenschaufel erfolgt etwas später ebenso wie die basale Verlängerung der Klebscheibe. Die Ontogenese zeigt, daß das primäre Element des Periplocaceen-Translators der obere Teil der Klebscheibe mit dem ihm dorsal ansitzenden basalen Stipesabschnitt ist.Innerhalb derAsclepiadaceae bilden die GattungenSecamone undFockea von den übrigen Gattungen abweichend gestaltete Translatoren aus, bei denen die Pollinien nicht an seitlichen Caudiculae, sondern an einer dorsalen Platte befestigt sind. Der morphologische Vergleich ergibt, daß die Translatoren beider Gattungen ausgeprägt plesiomorphe Züge tragen und als Zwischenformen zwischen dem Periplocaceen-Translator und dem elaborierten Asclepiadaceen-Translator angesehen werden können. Der Translator vonSecamone läßt sich mit dem primären Scutellum-Stipes-Komplex derPeriplocaceae homologisieren. Der Klemmkörper selbst entspricht einem nach vorn zusammengeschlagenen Scutellum, der die Pollinien tragende Dorsalfortsatz dem basalen Stipes.Für diese Homologisierung spricht nicht nur die relative Lage der Elemente der verglichenen Translatoren, sondern insbesondere ihre in den Grundzügen übereinstimmende Ontogenese. Von ausschlaggebender Bedeutung ist dabei die von der Schmalseite her erfolgende Bildung des Scutellums, die in der von der Basis aus fortschreitenden Anlegung desSecamone-Klemmkörpers ihre Entsprechung findet.Der Translator der GattungFockea stellt im Sinn des Stetigkeitskriteriums eine Zwischenform zwischen denSecamoneae und den elaboriertenAsclepiadaceae dar. Ohne diese Zwischenform wäre der morphologische Zusammenhang der Translatoren beider Taxa kaum zu belegen.Fockea stimmt mitSecamone in der Ausbildung des dorsalen Fortsatzes zur Pollinienanheftung wie auch in der Anheftung des adulten Translators an die Leitschiene mit Hilfe basaler Klebflächen überein. Eine Weiterentwicklung zeigtFockea in der Verlängerung des Translators parallel zur Griffelkopfoberfläche, die mit einer anfänglichen neuen Bodenbildung einhergeht. Diese neue Längsachse wie auch eine ausgeprägte Bodenstruktur sind charakteristisch für die Translatoren höher entwickelter Asclepiadaceen. Als auffälligstes Unterscheidungsmerkmal kommt bei ihnen die Befestigung der Pollinien mit Hilfe zweier Caudiculae hinzu. Aufgrund fehlender Zwischenstufen ist die phylogenetische Herkunft der Caudiculae bisher nur hypothetisch lösbar. Eine Ableitung von basalen Klebflächen wie etwa beiFockea wird für wahrscheinlich gehalten.Damit sind im elaborierten Asclepiadaceen-Translator nur noch Andeutungen von Elementen zu finden, die mit dem Translator vonSecamone homologisiert werden könnten. Nach den vorliegenden Untersuchungen erscheint der Entwicklungsschritt von der Periplocaceen-Stufe zum einfachsten Klemmkörper derAsclepiadaceae beiSecamone wesentlich geringer als die Evolution des Translators innerhalb derAsclepiadaceae selbst.

     

The turnover of milkweed pollinia on bumble bees,and implications for outcrossing

Summary Bumble bees (Bombus spp.) picked up a pollen packet (pollinium) of milkweed (Asclepias syriaca L.) every 2–5 h, usually more rapidly on their feet than on their mouthparts. Pollinia were retained an average of slightly over one day on the mouthparts and one-quarter day on the feet. This long retention period enhances the possibility of outcrossing in this largely or completely self-incompatible species. Although many more pollinia were carried on the feet, the longer retention of those on the mouthparts resulted in their collective outcrossing potential exceeding that of pollinia on the feet. Once pollinia became attached to the bees (via their attachment mechanisms, the corpuscula), others often attached to the translator arms (connecting corpusculum with pollinia) of the first pollinia. Long strings of pollinia and corpuscula often resulted, but they frequently were shed down to a single corpusculum, which was usually retained for long periods (estimated retention time=12–43 days). During the middle and latter part of the flowering period about two-thirds of the bees' feet carried only a single corpusculum. Feet in this condition picked up extremely few pollinia and thus were largely unavailable as sites for pollen transfer. Having several single corpuscula on the feet greatly lowered the number of pollinia carried by a bee over the season. Single corpuscula occurred much less frequently on the mouthparts and were shed over 25 times as rapidly as those on the feet. Many more bumble bees moved between clones than did other possible diurnal pollinators, largely a result of being several times more abundant than all other visitors combined. This factor, plus their heavy pollinia load, suggests that they were the most important diurnal pollinators of these clones.     

Temporal variation in pollinarium size after its removal in species of Bulbophyllum: A different mechanism preventing self-pollination in Orchidaceae

InBulbophyllum involutum andB. ipanemense (Orchidaceae), two closely related species, shortly after removal the pollinarium has a diameter of approximately twice that of the entrance of the stigmatic cavity, requiring a mean time of 105 to 135 minutes to shrink in width and allow pollination. Because the pollinators of these species remain for some minutes in the same flower after removing the pollinia this mechanism, previously unknown inOrchidaceae, may be very important in preventing self-pollination. This mechanism does not occur inB. weddellii, and the pollinator does not remain in the flower after removing the pollinia. The smaller diameter of the stigmatic cavity inB. involutum reduces by 50% the chances of interspecific pollination withB. weddellii, and interspecific crossing is strictly unidirectional. This is important in maintaining isolation between these sympatric species, which share the same pollinators and have synchronized flowering.     

Studies on the early floral development inCleomoideae (Capparaceae) with emphasis on the androecial development

Floral ontogeny ofCleome spinosa, Cleome violacea andPolanisia dodecandra subsp.trachysperma was studied in the context of the question whether the fascicled androecium ofReseda andCapparis (with fused fascicles) or the 2 + 4-pattern of theBrassicaceae is primitive in theCapparales. InPolanisia dodecandra, the 9–18 stamens show unidirectional initiation from the adaxial side toward the abaxial side of the flower. InCleome violacea, the six stamens also are formed in an unidirectional order, but development starts abaxially and a zigzag-like pattern is superimposed. InCleome spinosa, two stamen primordia in transversal (lateral) position are followed by four stamens which arise on a somewhat higher level in two pairs in front of the median sepals. It is assumed that the evolutionary steps in the androecial development proceed fromReseda viaCapparis andPolanisia/Cleome toBrassicaceae. This interpretation is supported byrbcL-studies (Chase & al. 1993,Rodman & al. 1993).Dedicated to emer. Univ.-Prof. DrFriedrich Ehrendorfer on the occasion of his 70th birthday     

Variation of Giemsa C-band and fluorochrome banded karyotypes,and relationships inAllium subgen.Molium

The somatic karyotypes of 10 taxa belonging toAllium subgen.Molium (Liliaceae) from the Mediterranean area have been investigated using Giemsa C-band and fluorochrome (Hoechst, Quinacrine) banding techniques. A wide range of banding patterns has been revealed. InAllium moly (2n = 14),A. oreophilum (2n = 16) andA. paradoxum (2n = 16) C-banding is restricted to a region on each side of the nucleolar organisers and the satellites show reduced fluorescence with fluorochromes. The satellites are also C-banded and with reduced fluorescence inA. triquetrum (2n = 18), but two other chromosome pairs also have telomeric bands which are not distinguished by fluorochrome treatment. InA. erdelii (2n = 16) 4 pairs of metacentric chromosomes have telomeric C-bands while 2 pairs of telocentric chromosomes have centromeric C-banding. InA. subhirsutum (2n = 14),A. neapolitanum (2n = 28),A. trifoliatum subsp.hirsutum (2n = 14) andA. trifoliatum subsp.trifoliatum (2n = 21) chromosomes with long centromeres, consisting of a centromere and nucleolar organiser are positively C-banded on each side of the constriction. InA. subhirsutum banding is confined to the pair of chromosomes with this feature, whereas inA. neapolitanum one additional chromosome pair has telomeric bands and inA. trifoliatum there are varying numbers of chromosomes with centromeric and telomeric bands, depending on the subspecies.A. zebdanense (2n = 18) shows no C-bands. The banding patterns in this subgenus are compared with those recorded for otherAllium species and with the sectional divisions in the genus. Evidence from the banding patterns for allopolyploidy inA. trifoliatum subsp.trifoliatum andA. neapolitanum is discussed.     

Allomyces neo-moniliformis gametogenesis

Summary InAllomyces neo-moniliformis meiosis takes place during resting sporangium germination. The meiospores are characteristically binucleate and biflagellate as described byEmerson (1938) andTeter (1944). A variation in the number of nuclei and flagella per meiospore from two is correlated with germination of the resting sporangia under reduced oxygen tension. The meiospores are extremely poor swimmers and are typically amoeboid. At encystment the gamma bodies of the cell are mobilized and appear involved in cyst wall synthesis. A single mitotic division of each nucleus gives rise to four nuclei. Gamete cleavage is as described for spore cleavage inBlastocladiella (Lessie andLovett 1968). The assembly of the nuclear cap and side body complex of the spore are extremely late processes in gametogenesis. The gametes are released when the single papilla dissolves. The gametes fuse in pairs and after zygote formation the cell is uninucleate with two flagella. The biflagellate zygote is an active swimming cell. The presence of homothallism or hetero-thallism inA. neo-moniliformis is discussed.     

New records of the carabidicolous laboulbeniales (Ascomycetes) of Japan

Six carabidicolous species of Laboulbeniales are reported as new for the Japanese mycoflora. They areDixomyces stomonaxi, Laboulbenia picardii, L. tenera, L. slackensis, L. aristata andL. kwangjuensis. Two forms ofD. stomonaxi are distinguished, one form of which resemblesD. nigromarginatus. A remarkable ornamentation consisting of a coillike pattern occurs on the receptacle ofL. picardii. InL. tenera andL. slackensis, the outer appendage has somewhat constricted, blackened septa near the base.Laboulbenia aristata andL. kwangjuensis have spirally arranged outer wall cells. Antheridia were observed inL. tenera, L. slackensis, L. aristata andL. kwangjuensis.     

Seed Epidermis Development and Histochemistry in Solanum melongena L. and S. violaceum Ort.

Structure, development and histochemistry of the seed epidermiswere studied inSolanum melongena L. andS. violaceum Ort. usinglight and scanning electron microscopy. The epidermal cellsat the endosperm mother cell stage of ovule development hadthickened outer periclinal walls, consisting of two layers,a thin inner layer, and a thick outer layer. The latter whichstained positively for pectic substances became further thickenedduring the course of seed development; more so inS. melongena.The inner layer of the outer periclinal wall also was thickenedby depositions of cellulose but remained comparatively thin.The development of the inner periclinal and anticlinal wallstook place by the uneven deposition of concentric layers. Thesesecondary wall thickenings which appeared as pyramids in transversesection stained for cellulose, lignin and pectin. Further unevensecondary thickenings near the outer part of the anticlinalwalls resulted in the formation of projections which were hair-or ribbon-like in appearance. InS. melongena, these projectionsprogressed only a short distance from the anticlinal wall. InS.violaceum, on the other hand, they grew much longer formingstriations on the inside of the outer periclinal wall. InS.melongena, partial removal of the outer periclinal wall by enzymeetching exposed to surface view a beaded appearance of the cellboundaries. Complete erosion of the outer periclinal wall revealedthe hair-like projections of the underlying anticlinal walls.InS. violaceum, enzyme treatment exposed the striations whichformed bridge-like structures over the curves in the anticlinalwalls. Solanum melongena ; Solanum violaceum; seed epidermis; seed structure; seed development; cell wall histochemistry; cell wall projections; cell wall striations     

Wood anatomy of theSarraceniaceae; ecological and evolutionary implications

The wood of theSarraceniaceae has a considerable number of primitive features including scalariform perforation plates, long and oblique end walls, scalariform lateral wall pitting, solitary vessels, tracheids with scalariform pitting, and diffuse axial parenchyma. Vessel elements in the genusHeliamphora have the greatest number of primitive features, while vessel elements inDarlingtonia andSarracenia appear to have modifications relating to temperate climates. The wood anatomy suggests thatHeliamphora is growing in a habitat more similar to the original habitat for the family thanDarlingtonia andSarracenia. The wood of theSarraceniaceae is similar to the wood of theTheales.     

The ontogeny of pollinia and elastoviscin in the anther of Doritis pulcherrima (Orchidaceae)

Wolter, M., Seuffert, C. & Schill, R. 1988. The ontogeny of pollinia and elastoviscin in the anther of Doritis pulscherrima (Orchidaceae). - Nord. J. Bot. 8: 77–88. Copenhagen. ISSN 0107–055X.
The ontogeny of pollinia of Doritis pulcherrima shows some features not commonly found in other angiosperms. The cell wails of the mature pollinium are almost identical with those of the pollen mother cells, and a callose envelope is lacking during the formation of sporopollenin. This fact may be related to the absence of a clear substructure of the exine and its low resistance to acetolysis. An exine is found only at the outer tetrads of the pollinium. The pollen grains in the tetrads possess irregularly spaced cell walls. The ontogeny of elastoviscin (specialized pollenkitt) begins in the ground cytoplasm of some tapetal cells, where small droplets are surrounded by myelin-like structures. Before anthesis the droplets fuse, flow into the loculus, and attach the pollinia to the stipe, thus forming a pollinarium. In the region in front of the elastoviscin-producing cells, some cells with specialized helical wall thickenings develop.     

Comparative structure of ovules and seeds inRafflesiaceae

The genera of theRafflesiaceae show a marked diversity in the structure of their ovules and seeds. Evolutionary trends are recognizable in ovule orientation and number of integuments. A change from anatropous ovules inApodantheae andMitrastemoideae towards incomplete anatropy inRafflesieae and orthotropy inCytineae occurs, next to a change from bitegmic ovules inApodantheae towards unitegmy with rudimentary outer integuments inRafflesieae andCytineae and full unitegmy inMitrastemoideae.—The differences in ovule structure are clearly reflected in the seeds. The seeds are essentially exotegmic, have very small embryos and an oily endosperm.—Seed structure strongly confirms the existing subfamilial classification and supports additional arguments for the generic status ofApodanthes. It does not support a separate status of the genusBerlinianche. InRafflesiaceae, seed micromorphology is only of limited use at the species level. As far as known seed dispersal is endo- or exozoochorous in all genera.     

Sectile pollinia and relationships in theOrchidaceae

The sectile state, in which pollinia are subdivided into units larger than tetrads, is one of several systematically important orchid pollinium characters. We recognize two types of sectile pollinia, based upon form and arrangement of massulae. One type is found primarily in members ofOrchidoideae-Spiranthoideae, while the other characterizes epidendroid genera. Additional characters that have systematic potential are massula dimorphism, caudicle composition, and degree of tetrad packing. Hollow sectile pollinia uniteArethusa andCalopogon; Stereosandra andEpipogium have dissimilar pollina, but other characters support a close relationship between them. The sectile condition is interpreted as resulting from a shift in the timing of cohesion of developing pollen cells.     

Ultrastructural aspects of leaves inFestuca ovina andPoa sphondylodes (C-3 Poaceae)

Structural aspects of the leaves of two common festucoids,Festuca ovina andPoa sphondylodes, have been examined employing the electron microscopy. The nature of vascular bundles and of sheaths that surround vascular tissues was discussed in the study. The festucoids exhibited a non-Kranz C-3 anatomy with more than four mesophyll cells separating the bundle sheaths of a leaf blade. Vascular tissues in theseFestuca andPoa leaves were surrounded by a double sheath: an inner distinct mestome sheath (MST) and an outer indistinctive layer of parenchymatous bundle sheath (PBS) cells. The PBS cells were much larger than the MST and had thin walls. The MST cells were relatively small and rectangular inP. sphondylodes and more or less hexangular in transverse sections ofF. ovina. InP. sphondylodes, MST had conspicuously thickened inner tangential walls with asymmetrically uninterrupted suberized lamellae in radial and tangential walls. In most differentiated MST cells, all walls were highly suberized. During suberin deposition, MST cells were quite vacuolated and most of the cytoplasm was present as a thin peripheral layer. However, MST walls inF. ovina revealed very thin suberized lamellae with translucent striations. No chloroplasts were detected inP. sphondylodes, whereas the MST inF. ovina contained small chloroplasts. Plasmodesmata were well developed in the primary pit fields of walls between MST and vascular cells, and between adjacent MST cells. Plasmodesmata were less frequent in the walls between the inner and outer sheath cells. Suberized lamellae were totally absent from the PBS cell walls in all veins. External to the PBS, the mesophyll comprised thin walled cells with abundant intercellular spaces. Peripherally arranged chloroplasts in the mesophyll were numerous and often larger than those of PBS and MST cells. Characteristics associated with C-3 and other ultrastructural features were also discussed in the study.     

Studies on systematic embryology inScilla (Hyacinthaceae)

Scilla persica and 5 species of the so-calledS. hohenackeri group, namely,S. furseorum, S. puschkinioides, S. vvedenskyi, S. hohenackeri, andS. greilhuberi, have been investigated embryologically with special reference to embryo sac and endosperm development.Polygonum-type embryo sac development was stated inS. puschkinioides andS. greilhuberi. 8-nucleate, normally structured embryo sacs, which could not be specified further due to sparse availability of the material, were stated inS. furseorum, S. vvedenskyi, andS. hohenackeri. InS. persica the embryo sac develops according to the bisporicAllium-type. In most species endosperm development was stated to be nuclear, exceptS. hohenackeri, where the type could not be specified. Other traits of possible taxonomic significance are the number of layers in the outer integument, which is mostly 4, or 5–6 inS. furseorum, and the occurrence of polyploid versus haploid and early degenerating antipodal nuclei, the latter occurring only inS. persica andS. furseorum. These embryological characters may be useful for assessing taxonomic relationship of the present species with other allied groups withinScilla, in particular, theS. siberica alliance,S. messeniaca, and theS. bifolia alliance. TheAllium-type embryo sac, which occurs inS. persica, is also characteristic for theS. siberica alliance, and may be a common derived character. Lack of antipodal polyploidization, as characteristic forS. persica andS. furseorum, occurs also in theS. siberica alliance, and is perhaps another common derived trait indicating phylogenetic relationship. Nuclear endosperm development is more frequent in spring-flowering squills than helobial development, which has previously been stated inS. messeniaca, some species of theS. siberica alliance, and inS. litardierei. While helobial endosperm may be primitive forHyacinthaceae in general, it may, by reversal, also occur as a derived character, at least in some species of theS. siberica alliance.     

The apical pore fibrous complex: a new cytological feature of some dinoflagellates

Summary A system of striated fibers has been discovered immediately beneath the apical pore of the dinoflagellatesHeterocapsa pygmeae andScrippsiella sweeneyae. The striated fibers radiate from a distinct fibrous ring lying beneath the apical pore plate. InScrippsiella, the fibrous ring is the outer edge of a fibrous apical disk that lies beneath the entire apical pore plate. The striated apical fibers radiate posteriorly and remain closely appressed to the sub-thecal membrane. The striated fibers terminate midway between the apical pore and the cingulum and are associated with sub-thecal microtubules inHeterocapsa.     

The tribal position of Fockea and Cibirhiza (Apocynaceae: Asclepiadoideae): evidence from pollinium structure and cpDNA sequence data

The pollinium morphology of the two members of the Asclepiadoideae, tribe Fockeeae, Fockea Endl. and Cibirhiza Bruyns, has been studied in detail and compared with that of eight genera of Marsdenieae, the tribe in which Fockea and Cibirhiza were previously accommodated and thus their putative closest relatives, as well as nine genera of Asclepiadeae. Both Fockea and Cibirhiza have several morphological characteristics in common, the most important of which is the absence of well-developed caudicula, which distinguishes them from all other genera of Asclepiadoideae known. The pollinium structure of these two genera, however, differs significantly. Whereas the pollinium of Cibirhiza consists of single pollen grains and is covered by a pollinium wall, as is typical for other Asclepiadoideae, the pollinium of Fockea consists of tetrads and is not covered by a pollinium wall, a condition otherwise typical of Secamonoideae. Fockea, however, has only two pollinia per anther, as does Cibirhiza and all other Asclepiadoideae, whereas the Secamonoideae have four pollinia per anther. Sequence data from two intergenic spacers, trnT-L and trnL-F and the trnL intron of cpDNA was analyzed. The ingroup included three species of Fockea and one species of Cibirhiza. The outgroup taxa consisted of three representatives each of Periplocoideae, and Secamonoideae and 24 species of Asclepiadoideae, including representatives of all tribes, of which eight genera belong to Marsdenieae, as outgroups. The results of the DNA analysis provide strong support for Fockeeae as a monophyletic tribe, distinct from Marsdenieae and, to the rest of the Asclepiadoideae. With the exception of pollen data, all morphological and molecular evidence clearly support recognition of the tribe Fockeeae. The occurrence of two such significantly different types of pollinia structure – characters elsewhere in the family used to distinguish subfamilies – within the small tribe Fockeeae was unexpected, and can perhaps best be understood as yet another attestment to the basal position of the Fockeeae in the nascence of the Asclepiadoideae.     

Floral development and systematics ofCistaceae

The floral development of representatives of six genera ofCistaceae has been studied. Calyx development involves the formation of a ring primordium in several taxa. Androecium development in species with intermediate or higher stamen numbers starts with the formation of a ring meristem on which the stamens are initiated in a centrifugal direction. In many taxa five alternipetalous leading stamen primordia can be observed. In the apetalous (cleistogamous) flowers ofTuberaria inconspicua androecium development appears to be unordered; this is probably due to the lack of petals. InLechea intermedia (also cleistogamous) the corolla is trimerous and three complex stamen primordia are produced, which give rise either to one or three stamens. Relationships withinCistaceae are discussed. Floral development inCistaceae is compared with that in otherMalvanae. Among the eight families ofMalvanae from which information on floral development is availableCochlospermaceae andBixaceae exhibit the greatest similarities toCistaceae. InCistaceae the leading stamen primordia are alternipetalous. InBixa the same condition seems to be present. InMalvales s. str. mostTiliaceae also show earliest stamen initiation in alternipetalous sectors, whereas the stamens of the innermost alternipetalous position are retarded early or even suppressed inSterculiaceae, Bombacaceae, andMalvaceae. WithinMalvales s. str. the diversity of androecial developmental patterns seems to decrease inBombacaceae andMalvaceae due to increasing synorganization in the mature androecium. The derivation of polyandry inMalvanae from diplo- or obdiplostemony is discussed by comparison with the sister clades ofMalvanae as shown in recentrbcL studies (i.e.Sapindales, Rutales, the glucosinolate producing clade, andMyrtales).     

Embryology ofMalaxis saprophyta,with comments on the systematic position ofMalaxis (Orchidaceae)

InMalaxis saprophyta, anther wall development corresponds to the Monocotyledonous type. The uninucleate tapetum is of secretory type and the endothecium develops U- and V-shaped thickenings on the inner tangential and radial walls. Cytokinesis is simultaneous; tetrahedral, isobilateral and T-shaped tetrads are formed which are compactly aggregated in pollinia. At anthesis the microspore tetrads are 2-celled. The ovule is anatropous, bitegmic and both integuments are dermal in origin. A single hypodermal cell develops directly into a megaspore mother cell. Embryo sac development is predominantly monosporic and less often bisporic. Irrespective of the type of development, the mature embryo sac is 6-nucleate. Although double fertilization occurs, the primary endosperm nucleus degenerates. Embryogeny is of the Onagrad type. The mature embryo lacks differentiation into cotyledon, plumule and radicle. The reticulate seed coat is formed entirely by the outer layer of outer integument. There are three sterile and three fertile valves in the ovary. Although initially parenchymatous, the entire three sterile valves in the ovary and the upper half of the three fertile valves become sclerified after fertilization. The embryological characters support the disputed systematic position ofMalaxis within subtribeMalaxidinae ofEpidendreae.     

Studies on the embryology and gynoecium structures inDrimys winteri (Winteraceae) and someAnnonaceae

Drimys winteri (Winteraceae) and 11 species ofAnnonaceae, namelyAnnona montana, Artabotrys hexapetalus, Bocagea sp.,Papualthia sp.,Polyalthia nitidissima, Tetrameranthus umbellatus, T. duckei, Uvaria sp.,Xylopia malayana, X. aromatica, andX. emarginata, were investigated embryologically with special reference to development of ovule and embryo sac. The ovules are anatropous, crassinucellate and in most taxa bitegmic. The inner integument is of epidermal origin. TheAnnonaceae investigated have a multi-layered, later vascularized outer integument with most probably subepidermal initiation. In contrast,Drimys winteri has a three-layered, non-vascularized outer integument of epidermal origin. The annonaceous genusTetrameranthus (T. umbellatus andT. duckei) possesses a middle integument between the inner and the outer one, stated here for the first time in a neotropic representative ofAnnonaceae. Within the angiosperms this feature occurs inAnnonaceae only. The embryological characters are rather homogeneous. Differences between the species investigated are found in, e.g. the number of cell layers in the inner integument, which is commonly two inAnnonaceae as compared to three inDrimys winteri, the presence or absence of a hypostase, the number of layers in the nucellar epidermis, great differences in size of ovules, and the species-specific pattern of tannin deposition in the ovules. In the species so far investigated the embryo sacs develop according to thePolygonum-type. InXylopia malayana andBocagea sp. in addition the carpels were investigated. They are conduplicate. InXylopia malayana the free carpels are united by an extragynoecial compitum, inBocagea sp. each stigma produces its isolated mucilage cap. The results obtained from the investigated taxa are discussed and compared with published data on embryology and gynoecium structure in other annonaceous and winteraceous taxa.     

Karyosystematics and evolution of AustralianAnnonaceae as compared withEupomatiaceae,Himantandraceae, andAustrobaileyaceae

Chromosome counts are presented for 12 genera and 20 species of AustralianAnnonaceae (all diploid with 2n = 16 or 18; Table 1) and two species ofEupomatiaceae (2n = 20, partly from Papua New Guinea). Detailed studies on interphase nuclear structure, condensing behaviour of chromosomes, and fluorochrome and Giemsa C-banding patterns also includeHimantandraceae (Galbulimima) andAustrobaileyaceae. — Eupomatiaceae completely correspond withAnnonaceae karyologically, their base number 2n = 20 is interpreted to have evolved from 2n = 18 by ascending dysploidy from common ancestors.Eupomatia laurina andE. benettii differ in DNA and constitutive heterochromatin (hc) quantity; their evolution from high to low DNA content probably corresponds to general progressions inMagnoliidae. Austrobaileya has nuclei of the presumably primitive Tetrameranthus type which is closely related to that ofGalbulimima and several other primitive taxa inMagnoliidae. Karyomorphology and other characters support the maintainance of two main branches within theMagnoliidae, Laurales andMagnoliales, withAustrobaileya probably intermediate; theWinteraceae appear more remote.—InAnnonaceae the reestablishment ofAncana is underlined by its chromosome number (2n = 18) the unexpected and specialized disulcate pollen, and various morphological characters which point to a close alliance with the Australian endemic generaFitzalania andHaplostichanthus (also disulcate) and the American genus pairSapranthus/Desmopsis; they are united in the provisionalSapranthus tribe, with a more distant position toFissistigma s. str. (2n = 16). AustralianAnnonaceae exhibit a high generic and a low species diversity; they can be considered as an ± old and partly impoverished outpost of the family with phytogeographical relationships to Asia, Africa and America.—On the base of field observations three main types of floral development inAnnonaceae are proposed, the most elaborated one found in the fly pollinated genusPseuduvaria. The growth form change from shrubs to lianas during the ontogeny ofDesmos andMelodorum, the vegetative propagation of anAncana species and the ecological and evolutionary patterns of the taxa investigated are discussed.